contaminationX

contaminationX estimates present-day human contamination in ancient male DNA samples by applying a maximum-likelihood model to low-depth X-chromosome sequencing data to distinguish endogenous ancient molecules from modern contaminants.

Key Features:

• X-chromosome maximum likelihood: Uses a maximum likelihood approach applied to X-chromosome sequence data from male individuals to infer contamination levels.
• Low-depth sequencing compatibility: Optimized for low-depth nuclear datasets and reports accurate estimates down to ~0.5× X-chromosome coverage when contamination is below 25%.
• Performance in challenging scenarios: Maintains accuracy under closely related target/contaminant populations and elevated sequencing error rates, as demonstrated by simulations.
• Efficiency: Computational runtime is reported as under 5 minutes for typical analyses.
• Implementation: Provided implementations in C++ and R.

Scientific Applications:

• Ancient human DNA studies: Provides contamination estimates to support authenticity assessments of endogenous ancient DNA in male samples.
• Analyses with closely related populations: Applicable when target and contaminant populations are genetically similar, where other methods may underestimate contamination.
• Low-coverage and error-prone datasets: Suited for studies with low sequencing depth or elevated error rates common in aDNA research.

Methodology:

Applies a maximum likelihood method to low-depth X-chromosome sequencing data from male individuals; performance and accuracy were evaluated using extensive simulations.